The goal of breast conservation surgery is local control through complete tumor removal with minimal normal tissue excision. However, in about one-third of lumpectomies, specimen margins contain cancer cells, strongly suggesting the presence of residual cancer in the breast. These residual cells are the presumed etiology of local recurrence, and are typically treated with local therapy (e.g., re-excision or radiation) or systemic therapy. Intraoperative ultrasound has been used to reduce the incidence of positive lumpectomy margins. Since ultrasound does not typically detect in situ cancer (which is associated with positive lumpectomy margins in 40% of cases), intraoperative ultrasound is not likely to completely solve the problem of positive lumpectomy margins. We propose the use of image-guided surgery in which a novel high resolution dedicated breast positron emission tomography ("PET") scanner is used to define the location and extent of cancer prior to surgery. Published pilot data suggest that this high performance scanner (2.5 mm full-width half-maximum spatial resolution) can accurately define the size of breast cancers. A prospectively planned controlled clinical trial is proposed, involving 264 women scheduled for lumpectomy who are randomized into two groups of equal size. In the test group, subjects will have high resolution PET scans of the affected breast prior to surgery, with guide wires or markers placed as needed to specify the cancer extent and location. Follow-up reviews of pathology records will determine the fraction of patients with positive lumpectomy margins. Members of the control group of subjects will not receive PET scans, and will have their records reviewed as in the test group. Statistical considerations suggest that we are likely to observe (with 80% power) whether the percentage of patients with lumpectomy margins decreases by a factor of two (i.e. from 30% to 15%) when the results from perioperative PET scans are applied. Matching private funds are already committed for all phases of this SBIR project.